HP StorageWorks 7110 - Virtual Array, Руководство по обслуживанию, Страница: 58 / 164

58 Product Overview
possibility of data loss in the event of multiple disk failures. Although the
statistical advantage of this minimal but measurable, HP advises when ever
possible to keep an even number of disks in each redundancy group.
For optimum availability, it is recommended that you have an even number of
disks in each redundancy group. Because of the manner in which the array
stores data, an even number of disks reduces the possibility of data loss in the
event of multiple disk failures. Although the possibility of this is extremely low,
using an even number of disk reduces the risk even further.
Auto Rebuild Behavior
(Firmware version HP14 and greater)
When a disk fails and Auto Rebuild is enabled, the array always attempts to
rebuild the data on the failed disk. This will occur even if the array may not
have enough capacity to complete the rebuild. For example, if hot sparing has
been disabled, there may not be enough capacity available to complete a
rebuild.
The array first makes an attempt to rebuild any data that was stored in
RAID1+0. This data is more vulnerable to another disk failure than data stored
in RAID 5DP. The array will continue to perform the rebuild until there is no
longer any capacity available to continue. This situation may result in
diminished performance when new data is written to the array in this
condition. The performance impact increases with the number of disks in the
redundancy group.
To avoid this situation, it is recommended that in configurations with 15 or
more disks per redundancy group, that Auto Rebuild is disabled whenever hot
spare is disabled.